A new approach to determining the risk of lethal prostate cancer upon initial diagnosis is sought to be developed in this project. Our approach relies on utilizing emerging stimulated Raman spectroscopic microscopy (SRSM) to record data from tissues, first in 2D then 3D. The spectral data will be related to cell type and its chemical changes in a case-control cohort of patients designed to be the most difficult 70% of contemporary prostate cancers diagnosed today. While past methods have focused on epithelial cells, we will comprehensively profile all cell types as well as the extra-cellular matrx through this one measurement. The properties of each component of tissue will be used to characterize the disease. Cellular abundance and spectral properties will be used in rick models to predict the chance of recurrent prostate cancer. The hypothesis underlying this effort is that the molecular and spatial information provided can complement and widely improve prostate cancer prognosis. The predictions will be compared to those made by existing clinical tools - Kattan and CAPRA-S nomograms- as well as an emerging, but not yet established, tool based on a 17 gene signature. The unique aspects of the approach in this project is that it uses a newly developed technology (SRSM), makes practical a concept that has not been realized yet for disease prognosis (using the tumor microenvironment) and directly seeks to compare with existing gold standards. The instrumentation, computer algorithms and patient cohort availability are unique resources that are augmented by almost decade long collaboration between engineers and pathologists for this project. A proposed set of consensus meetings to discuss progress, guide development and develop a protocol will engage the larger prostate cancer community and can lead to a translatable protocol at the end of the project period.